Winds and high energy flares in GRO J

Transcription

1 Winds and high energy flares in GRO J Emrah Kalemci Sabancı University Mitch Begelman (JILA), Tom Maccarone (Texas Tech), Tolga Dincer (Yale), Charles Bailyn (Yale-NUS), T. D. Russell (Amsterdam), John Tomsick (Berkeley)

2 Motivation Reanalyzed multiwavelength data of 2005 outburst rise of GRO J with the initial motivation of investigating hysteresis in spectral state transitions. Led to surprising (re-)discoveries: Hard X-ray flares in the soft state Iron absorption features (from a wind) together with optically thin radio flare (jet).

3 OUTLINE Introduction GRO J in 2005 Hard X-ray flares in soft state Iron lines with radio jets Future look to some wind studies with the extp

4 Why multi-wavelength studies? GRO J , Migliari et al Steady compact jet: flat to inverted radio spectrum up to a break in the Near/mid infrared > Conical Jet. We need to disentangle jet from the companion+disk +corona emission.

5 Black hole states and transitions, the if high inclination multiwavelength view See Homan et al. 2016

6 GRO J , a special dataset Well determined distance and mass Very rich multiwavelength dataset Third source showing relativistic jets (Hjellming & Rupen 1995) First radio hard X-ray correlation (Harmon et al. 1995) First magnetically driven winds Miller et al (but also see Netzer 2006) First mid-ir detection of (jet / disk/ combination) with Spitzer (Migliari et al. 2006) Rich near-infrared behavior (Shidatsu et al. 2016, Uttley & Klein- Wolt 2015) Rich X-ray phenomenology (many references!) High frequency QPO pair (Strohmayer 2001), low frequency QPOs deep in the hard state together with compact jet (Dincer et al. 2014).

7 OIR Flux (mjy) Radio Flux (mjy) Ecut, Efold (kev) PLF, DBB Tin (kev) G rms (%) GRO J , an archival MW study Reanalyzed RXTE in X-rays, SMARTS in optical/nir, and VLA in radio during outburst rise. Modelled RXTE PCA+HEXTE spectra with phenomenological, as well as Comptonization models (eqpair). Characterized state transitions, using timing and spectral data. Published recently in MNRAS, Kalemci et al PLF DBB HS J band flux 4.86 GHz Radio flux HIMS SIMS SS US Efold Ecut Time (MJD days) a b c d e f

8 OIR Flux (mjy) Radio Flux (mjy) HXR Flux 2 HXR Flux G GRO J , 2005, Hard X-ray flares Kalemci et al a b c d Time (MJD days) Tavani et al. 1996

9 OIR Flux (mjy) Radio Flux (mjy) EQW (ev) Ref. Frac. Ref. Flux Eqp&Dpn Flux Winds with radio flare 10-8 a b c 0.5 Iron absorption line energies and equivalent widths consistent with what XMM-Newton detected in the ultrasoft state (Diaz-Trigo et al. 2007). Figure also shows the need for a high energy component C1 X1 X2 d e Time (MJD days)

10 What can be done with the extp? Ponti et al unexplored! Kalemci et al 2013 Wind detections are scattered with Chandra and XMM-Newton (and Athena and X-ray Surveyor) difficult to schedule. Fundamental questions: how do winds form and die? How are they related to jet formation and quenching? Are they always thermal in origin but could they sometimes be magnetic? To be able to monitor wind related structures in X-rays everyday with mid-resolution will answer some of these questions and provide important insight onto origin of winds.

11 Future with extp 1. Wind absorption features as observed in GRO J during outburst rise will be detected in < 100s with LAD on extp allowing precise determination of wind formation and its relation to X-ray spectrum and compact jets. Determination of line centroid % error. 2. Wind absorption features stronger than 5 ev EW will be detected during outburst decay in an intermediate state in less than 1 ks with extp LAD for a typical flux level of a few x 10-9 ergs cm -2 s -1. Such detections will allow relating winds to jet formation. comparison of 3ks extp LAD and 80 ks XMM-Newton 1ks, typical intermediate state during decay, 7 ev EW feature.

12 Conclusions The quality of GRO J data provides us a glimpse of what awaits us with the extp in terms of observatory science. Thanks to better energy resolution the extp will provide much better constraints on wind structures, and will allow us to find how winds and jets are related when they are (either of them) first forming.

13 Flux (mjy) a Multi-wavelength transition during rise HS HIMS US Frequency (Hz) Time (MJD days) Concentrate on radio evolution for this source as the near-infrared to optical dominated by strongly irradiated accretion disk. Radio is compact jet dominated in the HS and HIMS, and optically thin jet dominated in the US. The evolution indicates opening up of the jet angle as the radio spectral index drops.

14 What about outburst decays? RXTE photon index SMARTS J mag 4.8 GHz radio 8.4 GHz radio radio data from Corbel et al. 2013

15 All sources we investigated Similar rebrightinings, in other sources and outbursts with similar spectral index behavior! Natural to assume that the NIR brightening is jet related, but also see Veledina & Poutanen 2012 Poutanen & Veledina 2014 hot accretion flow model. Kalemci et al. 2013

16 Comparison of rise and decay sequence spectral index (Γ) as a proxy for strength of corona, formation of geometrically thick medium NIR and radio flux as a proxy for jet emission Radio spectral index as a proxy for jet collimation.

17 RISE DECAY

18 A new hysteresis Launch during decay Quenching during rise Γ Jet emission 5-10 days time (days) Γ 5-10 days Jet emission time (days) Wait B flux to accumulate B flux diffuses out, not replenished Kalemci et al. to be submitted

19 Extended future Future outlook II Prepare for Athena and extp with WINDS of change. Why don t we observe winds in the hard state? Surveys LSST, SKA + SMARTS / DAG NIR complete and excellent MW coverage of state transitions! DAG Telescope with possible MKID focal plane instrument

20 Elevated Disks and State Transitions Elevated disks are geometrically thicker, and accretes faster, and thermally and viscously stable. The result: MRI active lower zone produces thermal emission MRI active upper zone produces thermal + nonthermal Comptonized emission, winds, jets Upper vs. lower luminosity ratio determines state Lower dominates: soft state Upper dominates: hard state Both comparable: intermediate state (Begelman et al. 15)

21 Origin of the poloidal field? The secondary star? Advection/stochastic accumulation? Stochastic net magnetic field accumulation timescale: Begelman & Armitage 2014 with typical parameters, it takes days to obtain magnetic fields high enough to produce jets, it is the right time scale!

22 Magnetically arrested disks and field flipping Tchekhovskoy et al See also Li & Yan 2015 Igumenschev 2009 Dexter et al

23 Disk breaking. Could only explain BH outbursts with disk/spin misalignment, not valid for NSs and BHs but, May explain: Quick HIMS/SIMS transitions. Nixon & Salvesen 2013

24 Questions What determines the transition luminosity in the upper branch? Is the transition luminosity constant during the decay? When is the optically thin jet actually launched during the rise, how is it related to timing and spectral changes? What is the role of B field in state changes, and where is it created? How do you reconcile truncated disk models with iron line transition radius measurements? What is the role of boundary layer in launching jets in WDs and NSs? Can a single mechanism explain major properties of state transitions in BHs, WDs and NSs? Outlook Phenomenology is way ahead of theory Expecting big breakthrough in simulations. SKA + LOFTish + LSST + IR monitoring would open up new insights.

25 DISCUSSION and INTERPRETATION I Observing compact jet requires: 1. Transport of ordered magnetic field (Beckwith 2009, McKinney 2012), 2. Effective collimation, 3. Radiative processes (shocks, etc, Fragile et al. 2012) Optically thin radio emission: Outflows, in terms of wind/or not so collimated structures may occur as soon as we see the changes in X-ray spectral and temporal parameters, basically the corona cannot collimate outflows. bi-polar outflow, optically thin, not collimated, A Chandra proposal has been written to test this hypothesis.

26 Disk flux goes DISCUSSION down corona and scale INTERPRETATION height increase IIeffective transport of B field + effective collimation, all ingredients necessary for compact jet formation. BUT...

27 Spectral state transitions 3-79 kev NuSTAR spectra (King et al. 2014; Fuerst et al. 2015) Sudden transitions during an outburst from XTE J Soft state: ~1 kev thermal plus power-law Hard state: Γ ~ power-law 27 2/18/2017

28 What is the source of ordered B field? Outer disk, companion star? high B companion (Shuang-Lian and Yan 2015) Thick disk/thin disk boundary: Begelman & Armitage 2014 RP cosmic battery (Contopoulos & Kazanas 98)

29 Hysteresis and state transitions Begelman & Armitage 2014 Cao 2015, magnetic outflows increasing radial velocity, helping ADAF survive. Kylafis had similar ideas

30 Interpretation in terms of B field accumulation-transport models The magnetic field normally diffuses out in the soft state when the optically thick disk moves all the way to ISCO. During the hard state, magnetic field in the inner region is continuously replenished (perhaps by the mechanism described in Begelman & Armitage 2014 or PRCB). As the inner disk forms, the easy transport is hampered, and B field decreases, reducing collimation and decreasing radio/infrared flux, which is consistent with what we observe during the rise. This would happen as soon as disk starts to move in because the field is already present in the inner regions. While during the rise we need to wait until enough B field is accumulated. The prediction is then the softening should also lead to radio spectrum becoming optically thin.

31 X-ray emission from black holes Comptonized (power law) Reflected Corona (~100 kev) Disk (disk blackbody) Disk (0.1-1 kev) BH Phenomenological model: disk-blackbody (diskbb) + Comptonized hard X-rays (power-law)+ +reflection +iron line and edges Energy (kev)

32 Softening? sources for which the flux drops rapidly (and the index change rapidly) show softening.

33 For the jets already formed Simple scaling predicts break in the SED going to higher frequencies with Falcke & Biermann 1995, Heinz et al. 2003, Markoff et al. 2003, Falcke et al. 2004: increasing jet power. We should see radio before the NIR. MAXI J it is a failed outburst... The compact jet was always there. Russell 2013

34 All sources well covered 5 sources, 8 decays + not so well covered 2 sources, 4 decays ELF: kev Luminosity/L Edd Jets appear at around a few % L Edd The delay between index transition (spectral hardening) and compact jet appearing is 5-25 days... *For sources with no distance information 8 kpc used. For sources with no mass information 8M used (Ozel et al 2010, Kreidberg 2012)

35 Evolution of power-law vs disk luminosity Power law luminosity Power law flux increases rapidly with the timing transition, but often the compact jets are observed DAYS after the power law ELF peaks... DAYS!!!! dynamical timescales

36 Method For each X-ray observation, fit disk+power-law, or your favorite model Find rms amplitude of variability using power spectra, get QPO information (if possible) Plot parameters as a function of time Add multiwalength observation (IR, optical radio) Find transitions. Kalemci et al power-law spectral index NIR light curve disk flux power-law flux

37 TRANSITION SEQUENCE in OUTBURST RISE: 1. Softening and IR drop simultaneously compact jet turning off. 2. A transition in timing. 3. Complicated HIMS SIMS SS transitions fxpsd (rms/mean) J band flux density radio flux density Frequency (Hz)

38 Transition sequence during decay: 1. Timing transition Time (days)

39 2. Index transition, spectral hardening

40 3. Compact jet transition, OIR brightening Fit increase in near-infrared fluxes, or, first radio observation with flat to inverted spectrum How do we know near infrared rise correspond to compact jet formation?

41 H in 2009 Miller-Jones et al

42 Frequency (Hz) fxpsd (rms/mean) fxpsd (rms/mean) 2 SOFT fxpsd (rms/mean) 2 X-ray states and timing Frequency (Hz) Rms<15% Rms<5% SIMS Frequency (Hz) fxpsd (rms/mean) HIMS Frequency (Hz) fxpsd (rms/mean) Timing is a better proxy for determining intermediate states, Belloni et al. Rms>15-30% Rms>30% Frequency (Hz) HARD STATE